An increased risk of malignancy has been shown in multiple studies with thiopurines, but whether anti-TNF therapy increases the risk of malignancy is less evident. It also is unknown whether combination therapy with 2 immunosuppressant agents with different mechanisms of action increases the risk of malignancy compared with the use of either drug alone. In this pooled prospective analysis of randomized controlled trials (RCTs) and a long-term extension study for treatment of active CD with adalimumab, we found that adalimumab monotherapy was not associated with an increased risk of NMSC or other malignancies when compared with the general population. However, combination therapy with adalimumab and immunomodulators was associated with a 3-fold increased relative risk of malignancies other than NMSC and a 5-fold increased relative risk of NMSC compared with the general population, and similarly a 2.6- to 4-fold increased relative risk of these malignancies compared with patients receiving adalimumab monotherapy. In terms of absolute risk, combination therapy was associated with an increased incidence of malignancies of 0.6 per 100 person-years for both NMSC and other malignancies compared with adalimumab monotherapy.
Our finding that adalimumab monotherapy was not associated with an increased risk of malignancies other than NMSC is in keeping with 3 recent pooled analyses of anti-TNF use in autoimmune disease, including a pooled analysis of 10 placebo-controlled trials of infliximab therapy in IBD, a meta-analysis of 74 RCTs including more than 22,000 patients treated with anti-TNF therapy for a variety of autoimmune diseases (rheumatoid arthritis, psoriatic arthritis, psoriasis, IBD, ankylosing spondylitis, and other diseases), and a pooled analysis of 71 global clinical trials of adalimumab in more than 23,000 patients with various autoimmune diseases (rheumatoid arthritis, juvenile idiopathic arthritis, ankylosing spondylitis, psoriatic arthritis, psoriasis, and CD). All 3 of these analyses excluded NMSC from the outcome. In addition, several analyses assessing the risk of all-site malignancies (which included NMSC) with anti-TNF therapy in IBD and other autoimmune diseases reported that anti-TNF treatment was not associated with increased risk. Specifically, these analyses included a meta-analysis of placebo-controlled trials of various anti-TNF drugs in CD, data regarding infliximab safety from the TREAT registry, and a meta-analysis of 7 prospective observational studies of anti-TNF therapy in rheumatoid arthritis.
Our finding of a nearly 3-fold increased risk of malignancies other than NMSC with combination adalimumab and immunomodulator therapy compared with adalimumab monotherapy is novel. A recent pooled analysis of placebo-controlled trials of infliximab therapy in IBD showed that patients treated with infliximab (alone or with concomitant immunomodulators) or immunomodulators (alone or with concomitant infliximab) did not have an increased risk of malignancies other than NMSC compared with the general population; in addition, patients treated with immunomodulators (with or without concomitant infliximab) did not have a higher risk of malignancy than those treated without immunomodulators (with or without concomitant infliximab). However, in contrast to our analysis in which the risk of malignancy among patients receiving combination therapy were clearly different from those receiving adalimumab monotherapy, the analysis of patients treated with infliximab did not distinguish between patients receiving combination therapy and infliximab monotherapy and thus the relative risk of malignancy with combination therapy compared with infliximab monotherapy was not reported.
Prior studies showing an increased risk of lymphoma with thiopurines support the hypothesis that combination therapy would have a higher risk of malignancy than monotherapy. Recent data assessing the risk of combination therapy in IBD from a large meta-analysis and the prospective CESAME cohort, both of which showed a similar risk of lymphoma with combination therapy compared with thiopurine monotherapy, suggest that the risk of lymphoma with combination therapy is largely conferred by thiopurine use. However, the 2 lymphomas in patients receiving combination therapy in our analysis are not enough to fully explain the association we observed. Evidence for thiopurines and methotrexate to increase the risk of other cancers is strongest for NMSC, although there is growing evidence for other malignancies as well. A recent case-control study from a French referral center observed that long-term use of azathioprine was associated with a 2.2-fold increased risk of overall malignancies compared with nonusers of azathioprine. A large Danish cohort study, which included more than 5000 users of azathioprine, also found a significantly increased risk of overall malignancies with use of azathioprine (rate ratio, 1.41; 95% CI, 1.15–1.74). In addition, the prospective CESAME cohort reported that the incidence of overall malignancies in thiopurine users was twice as high as that in nonusers (2% vs 1%, P = .0016). Notably, many of the malignancies observed in these studies were similar to those seen in our analysis.
It is unlikely that detection bias by being in a clinical trial accounted for the association that we observed because the incidence rate of malignancy in the placebo group was very low (0.7 per 100 person-years), albeit with short follow-up time, and no increased risk of cancer was seen with adalimumab monotherapy. It is also unlikely that differences in disease severity between adalimumab monotherapy and combination therapy explained our results because all patients included were enrolled in a clinical trial and we controlled for disease duration, baseline CDAI, baseline corticosteroid use, and prior anti-TNF therapy in the adjusted analyses. In addition, patients receiving adalimumab monotherapy were more likely to have an increased baseline C-reactive protein concentration than those receiving combination therapy.
In our analysis, we observed that adalimumab monotherapy did not increase the risk of NMSC in patients with CD (SIR, 1.2; 95% CI, 0.4–2.8). This result is generally in keeping with several prior studies. Two recent nested case-control studies using large national administrative databases and controlling for the use of other immunosuppressive medications found conflicting results. The first of these studies showed a 2-fold increased risk of NMSC in CD with recent (in the past 90 days) or persistent (>365 days) use of anti-TNF therapy, whereas the second study did not observe an increased risk of NMSC in patients with IBD treated with anti-TNF medications. Three other studies addressing the risk of NMSC with anti-TNF therapy in autoimmune diseases have showed either no increased risk or a modestly increased risk.
Our finding that combination adalimumab and immunomodulator therapy was associated with a 3.5- to 4-fold increased risk of NMSC compared with adalimumab monotherapy is not surprising. It is likely that thiopurines independently increase the risk of NMSC in patients with IBD because 4 studies (3 nested case-control studies using large national administrative databases and the prospective CESAME cohort) have reported significant 2- to 6-fold increased risks with thiopurine use, in contrast to only 1 study using a large Dutch administrative database that did not show an increased risk. Two of these studies assessed the impact of combination therapy, one of which observed similar risks of NMSC with recent or persistent use of combination therapy compared with either thiopurine or anti-TNF therapy in CD, whereas the other found an increased risk in IBD with any exposure to combination therapy vs thiopurine or anti-TNF therapy alone. In the Safety of Biologic Therapy cohort, the risk of NMSC was lower among patients treated with anti-TNF therapy with or without concomitant immunomodulators than those receiving thiopurine monotherapy. Taken together, these data suggest that the risk of NMSC is greater with thiopurines than with anti-TNF therapy and that the increased risk of NMSC among patients receiving combination therapy is likely mostly attributable to the use of thiopurines.
Of note, no melanomas were seen in our analysis, despite a recent nested case-control study that showed a 2-fold increased risk of melanoma with anti-TNF therapy in CD, and studies in rheumatoid arthritis suggesting a 2-fold increased risk of melanoma with anti-TNF therapy. A variety of other malignancies were observed, some of which historically have been associated with the use of thiopurines or anti-TNF agents and others of which have not.
Overall, our analysis adds to a small but growing body of data on the risk of malignancy with anti-TNF monotherapy and combination therapy in autoimmune diseases and particularly IBD. Our finding that combination therapy increases the risk of collective malignancies other than NMSC in CD is novel and requires further exploration and confirmation, but may be owing to the effect of thiopurine use. The increased risk of NMSC with combination therapy in our analysis is quite possibly owing to the effect of thiopurine use.
If there truly is an increased risk of malignancy with combination therapy compared with anti-TNF monotherapy in CD, it must be put into the appropriate context. Although we found a relative risk of malignancy of 2.6–4 with combination therapy, the absolute excess risk of 0.6 events per 100 patient-years for both categories of malignancies translates to a number needed to harm of 167. To put this into context, one must consider the potential benefit to be derived from combination therapy in patients naive to both medications. In SONIC trial, infliximab plus azathioprine resulted in higher rates of corticosteroid-free remission at 26 weeks (number needed to treat, 9) and lower rates of any serious adverse events over 54 weeks. Whether similar benefits would be realized with combination adalimumab and azathioprine in patients with CD naive to both drugs is unknown. Also unknown is whether continuing failing immunomodulator therapy in patients newly started on anti-TNF therapy is beneficial. This scenario represents the combination therapy used in our analysis as well as the large RCTs of infliximab for IBD.
Our analysis had several notable strengths that distinguish it from prior studies of malignancy risk with anti-TNF therapy. We were able to combine both randomized clinical trial data with prospective observational data, thereby allowing us to benefit from the longer follow-up time of the extension study and prospective data collection. In addition, we used 2 types of comparator groups, the general population and CD patients with a different degree of immunosuppression, in an effort to capture the risk of malignancy with anti-TNF therapy more completely. Furthermore, we were able to analyze separately the risks of NMSC and other cancers because combining these 2 categories of malignancy may not yield the most accurate description of risk.
This analysis had several limitations as well. First, was the absence of an immunomodulator-only group. Second, we recognized that the general population of non-IBD patients was an imperfect comparator for IBD patients treated with 1–2 immunosuppressants. However, the SIRs are useful for clinicians to put the absolute and relative risks of malignancy in perspective, especially for patients receiving adalimumab monotherapy. Third, the median follow-up time was shorter than 2 years. As such, it is possible that we may have underestimated the relative risk of malignancy with combination therapy if this increased with duration of therapy. Likewise, we could have missed an association with adalimumab monotherapy if the latency period was much longer. Fourth, the use of immunomodulators was defined as baseline use only and was not adjusted for in a time-updating manner. Given the CESAME data that prior thiopurine exposure places patients at increased risk of NMSC, adjusting for prior use of immunomodulators (especially in the adalimumab monotherapy group) likely would have increased the risk difference that we observed, giving rise to an even higher relative risk for NMSC with combination therapy. Although the CESAME data found that prior thiopurine use did not increase the risk of lymphoma, it is unclear whether prior thiopurine use would influence the rate of other malignancies, and thus it is difficult to surmise whether adjusting for prior thiopurine use for the outcome of malignancies other than NMSC would alter the outcomes. Finally, data were not available to allow us to adjust for radiation exposure from imaging studies. Although it is possible that some patients receiving combination therapy had more severe disease and thus required more frequent imaging studies, which would have led to higher radiation exposure, it is doubtful that this difference in radiation exposure was enough to account for the entire magnitude of the associations seen. We were able to adjust for race, but not sunlight exposure. However, it is unlikely that this exposure would have differed between the adalimumab monotherapy and combination therapy groups and there is also no reason to suspect that IBD patients would be exposed to more sunlight than the general population.
In conclusion, with a median follow-up period of approximately 1.5 years, we found that patients with CD treated with adalimumab monotherapy did not have an increased relative risk of malignancy compared with the general population. In contrast, patients treated with combination adalimumab and immunomodulators had a 3- to 5-fold increased relative risk of malignancy compared with the general population and a 2.6- to 4-fold increased relative risk compared with patients treated with adalimumab monotherapy. These data suggest that the increased risk likely is attributable to the immunomodulator therapy. Importantly, for most patients, particularly younger patients, the absolute increase in risk of cancer is small and must be weighed against the potential benefits of combination therapy on the course of CD. Furthermore, these data are important when contemplating the role of thiopurine monotherapy for anti-TNF–naive patients with CD who are reluctant to pursue combination therapy.